High current during WFE with nRF52832

[eBrock]

I measure 184 uA of current during WFE. I've read in other posts that the nRF52832 should draw about 1.2 uA during WFE.

I've looked over the project settings many times and everything looks correct. I created a very basic firmware to investigate

this issue, which is posted below. Any help is greatly appreciated.

Nordic SDK:  15.3.0

SES v4.22

Hardware is nRF52 DK

Preprocessor definitions:

NO_FPU_ENABLE

FLOAT_ABI_SOFT

INITIALIZE_USER_SECTIONS

NO_VTOR_CONFIG

NRF52

NRF52832_XXAA

NRF52_PAN_74

Code:

#include <stdbool.h>
#include <stdint.h>
#include <string.h>
#include "nrf.h"
#include "nrf_delay.h"
#include "nrf_gpio.h"

#define LED_1 17

void clocks_start( void )
{
// Start HFCLK and wait for it to start.
 NRF_CLOCK->EVENTS_HFCLKSTARTED = 0;
 NRF_CLOCK->TASKS_HFCLKSTART = 1;
 while (NRF_CLOCK->EVENTS_HFCLKSTARTED == 0);
 nrf_delay_ms(1);
}

void gpio_init( void )
{
 nrf_gpio_cfg_output(LED_1);

}

int main(void)
{
 clocks_start();
 gpio_init();

 while (true)
 {
  nrf_delay_ms(2000);
  __WFE();

 }
}

[eBrock]

According to this table:  https://infocenter.nordicsemi.com/topic/com.nordic.infocenter.nrf52832.ps.v1.1/pmu.html?cp=3_2_0_16_0_0_2#unique_498092668

I would assume that the WFE task is able to stop/start the HFCLK automatically, but in my testing, that isn't happening.

I don't see any errata for this issue. If there is a device problem, I'd like to know, so I can work on an alternative strategy for low-power mode.

[Dmitry]

WFE is able to stop/start internal HFCLK (HFINT) automatically. If you remove clocks_start() from your code, device will be clocked from HFINT, and it will stop at WFE, assuming you have no running peripheral that needs HFCLK.

External oscillator (HFXO) is started and stopped explicitly with TASKS_HFCLKSTART/TASKS_HFCLKSTOP,  not by WFE.

[eBrock]

Ahh, OK, that makes sense now. Thanks for explanation.

My application will use the radio periodically, so I need the HFXO.